A milliarcsecond localization associates FRB 20190417A with a compact persistent radio source and an extreme magneto-ionic environment

Abstract

We report the milliarcsecond localization of a high (1379 pc/cc) dispersion measure (DM) repeating fast radio burst, FRB 20190417A. Combining European VLBI Network detections of five repeat bursts, we confirm the FRB's host to be a low-metallicity, star-forming dwarf galaxy at z = 0.12817, similar to the hosts of FRBs 20121102A, 20190520B and 20240114A. We also confirm that it is associated with a previously reported persistent radio source (PRS), which is compact on milliarcsecond scales. Visibility-domain model fitting constrains the transverse physical size of the PRS to < 23 pc and yields an integrated flux density of 191(39) microJy at 1.4 GHz. Though we do not find significant evidence for DM evolution, FRB 20190417A exhibits a time-variable rotation measure (RM) ranging between +3958(11) and +5061(24) rad/m2 over three years. We find no evidence for intervening galaxy clusters in the FRB's line-of-sight and place a conservative lower limit on the rest-frame host DM contribution of 1228 pc/cc (90% confidence) -- the largest known for any FRB so far. This system strengthens the emerging picture of a rare subclass of repeating FRBs with large and variable RMs, above-average host DMs, and luminous PRS counterparts in metal-poor dwarf galaxies. Our results suggest that these systems are the result of environmental selection, or a distinct engine for FRB emission.

Figures (6)

• Figure 1: EVN dirty map of FRB20190417A-PRS, as seen in the combined five epochs. A small bar at the top right of the image shows a representative 50-pc transverse extent, for scale. Contour levels start at two times the rms noise level of $12\rm\,\upmu Jy~beam^{-1}$ and increase by factors of $\sqrt{2}$. The dashed circle shows the 1-$\sigma$ VLA positional uncertainty of the PRS 20190417A-S1 ibik2024prs. The synthesized beam is represented by the ellipse at the bottom left corner; it has a major and minor axis of $32.7$ and $30.6\rm\,mas$, respectively, and a position angle of $-10.8\degr$.
• Figure 2: Top left:$0.5\arcsec \times 0.5\arcsec$ dirty image of five combined FRB20190417A bursts detected in our EVN-PRECISE observations. The solid contours are taken from Figure \ref{['fig:PRS']} and represent the EVN position of FRB20190417A-PRS, while the dashed circle shows the 1-$\sigma$ VLA positional uncertainty ibik2024prs. The synthesized beam is represented by the ellipse at the bottom left corner of the top left panel; it has a major and minor axis of $38.8$ and $26.1\rm\,mas$, respectively, and a position angle of $38.5\degr$. Top right: A zoomed-in image on the white square shown in the left panel. Bottom: CDF of the pixel values in a zoomed-out $2\arcsec \times 2\arcsec$ field. The black-dotted line shows the pixel value at the position of the EVN PRS centroid, while the gray-shaded region denotes the pixel values within $1\sigma$ of the nominal PRS position.
• Figure 3: The host galaxy of FRB20190417A in the context of other FRBs and energetic transients. The known FRB-PRS systems (listed in Table \ref{['tab:frbprsproperties']}) are shown as colored stars while other FRB hosts are shown as green stars and squares (for repeaters and non-repeaters, respectively). Left: Host galaxy apparent r-band magnitude versus redshift. Information for FRB hosts was taken from Gordon2024 and references therein. For comparison, we also show a set of SLSN-I (pentagons) and LGRB (circles) host galaxies from Lunnan2014 and Svensson2010. The host galaxies of the FRB-PRS systems are notable in that, with the exception of FRB 20201124A, they are all fainter than $0.1$ L$_*$ galaxies. Right: Galaxy stellar mass versus gas phase metallicity. Black lines represent the median, 68%, and 95% contours for SDSS star-forming galaxies from Tremonti2004. FRB host galaxy information is taken from Heintz2020Bhandari2022 and Bhardwaj2021, SLSN-I from Lunnan2014 and LGRB from Levesque2010. The metallicity for the SLSN-I and LGRBs was measured either using an O3N2 diagnostic Hirschauer2018 or an R23 diagnostic Kobulnicky2004. To compare our PP04 N2 measurement from Section \ref{['sec:spectrum']}, we convert to R23 using the equations from Kewley2008. The host galaxies of the FRB-PRS systems are again notable: they have low masses and metallicities, more comparable to those of SLSN-I and LGRB hosts.
• Figure 4: Figure 3 from niu2022r1twin updated to include $51$ FRBs with robust ($P_{\rm PATH}>0.9$) host associations and redshifts outrigger2025law2024. Galactic disk contributions are estimated from NE2001 with $\pm20\%$ uncertainties, along with an additional halo contribution of $25$--$80$ pc cm$^{-3}$. The red error bars on each extragalactic DM estimate represent a conservative full range uncertainty encompassing the Galactic disk and halo ranges. The expected median DM contribution of the IGM and the inner 1-$\sigma$ confidence interval are given by the orange line and shaded region, respectively. The known PRS-associated FRBs are shown as colored stars while other FRBs are shown as green stars and squares (for repeaters and non-repeaters, respectively). FRB 20190520B intersects an exceptional LoS lee2023flimflam; we include its estimated DM$_{\rm halos}$ contribution for reference.
• Figure 5: Temporal profiles (top) and dynamic spectra (bottom) of the eight FRB20190417A bursts detected with the Ef radio telescope. Each burst is coherently dedispersed to a DM of $1379.0\rm\,pc\,cm^{-3}$ and is shown with a frequency resolution of $1$ MHz; the time resolutions are shown in the top right corners of each panel. Horizontal white bands correspond to frequency channels that have been flagged due to RFI. The limits of the color map have been set to the $20^{\mathrm{th}}$ and $98^{\mathrm{th}}$ percentile of each dynamic spectrum. We note some of the wider bursts appear slightly over-dedispersed; future work will analyze all bursts at a higher time resolution to place tighter constraints on their DM.